Understanding eruptive style variations at calc-alkaline volcanoes: the 1913 eruption of Fuego de Colima volcano (Mexico)

The 1913 sub-Plinian eruption of Fuego de Colima volcano (Mexico) occurred after almost 100 years of effusive and (minor) Vulcanian explosive activity, which modulated dome growth and destruction. Dome extrusion persisted from 1869 to 1913. The transition to explosive eruption started on 17 January 1913, and it progressed in three phases: (1) opening, with the generation of block-and-ash flows, (2) vent clearing, with strong explosions that destroyed the summit dome and decompressed the magmatic system, and (3) sustained column (sub-Plinian fallout) with final collapse producing pyroclastic density currents. Because of this succession of events, the 1913 activity represents an excellent case-study for investigating the eruptive style changes at calc-alkaline volcanoes. We investigated the conditions that led to eruptive style transition from effusive (dome growth) to explosive (the final sub-Plinian fallout) through steady-state numerical simulations, using subsurface data and independently inferred (from volcanological data) mass discharge rates as constraints. Results show good matches for hybrid geometrical settings of the shallow conduit-feeding system (i.e., dyke developing into a shallower cylindrical conduit), and the magma chamber top at 6 km of depth. The fragmentation level was shallower than 2 km, as inferred from the lithics contained in the sub-Plinian fall deposits of Phase (3). The most likely solution is represented by a dyke having major axis between 200 and 2000 m and the minor axis of 40 m. The dyke-cylinder transition was set at a depth of 500 m, with a cylinder diameter of 40 m. It emerges that at least two main mechanisms may be responsible for the effusive to explosive transition that led to the Phase (3) of the 1913 eruption: (i) an increase in magma chamber overpressure (magmatic triggering) or (ii) decrease of lithostatic stress acting on the volcano (non-magmatic triggering). The former implies arrival into the magma chamber of a batch of fresh magma, which can have volume between 10 and 200 × 106 m3, depending on the values of magma and host rock compressibility. The latter requires decompression-induced emptying of at least the equivalent of 1000 m of the magma column to produce the necessary unloading of the conduit-feeding system. A sudden jerk in the lateral spreading of the Fuego de Colima cone would be a reliable mechanism for decompressing the upper conduit and driving fragmentation processes over a time period of few hours. The results are not conclusive on an internal (magma chamber overpressure), external (lowering of lithostatic load), or mixed (internal and external) trigger of the observed eruptive style transition. This work highlights how different processes can have non-linear cascade effects on close-to-equilibrium volcanic systems like Fuego de Colima volcano.